The present invention relates to a scanning optical microscope which disperses the fluorescence from a sample into a plurality of wavelength ranges and detects the fluorescence of each wavelength range.
The recent fluorescent observation often uses multiple dyes as well as a single dye. Since fluorescent dyeing is performed to permit cells or a specific target in an organ to be observed, each dyed portion should be detected as a clear color difference or a clear difference in fluorescent wavelength in the multiple dye observation. In this case, it is necessary to effectively remove the partial overlapping of the fluorescent wavelengths (crossover portion) in the detection. The fluorescent observation also demands a high contrast and high optical resolution. Confocal scanning laser microscopes satisfy those requirements and are becoming popular in researches in the field of biology.
Confocal scanning laser microscopes to which this invention relates and which can ensure fluorescent observation are disclosed in Jpn. Pat. Appln. Kokai Publication Nos. Hei 8-43739 and Hei 9-502269. Those microscopes use spectral resolving means like a prism or diffraction grating as fluorescence separation means for multiple dyes, and a slit for restricting the fluorescent wavelength range. This can ensure highly efficient detection of fluorescent rays from a multi-dyed sample without crossover while achieving the high contrast and high resolution of a confocal microscope.
The fluorescence from a sample is generally so weak that a photomultiplier is needed as a photosensor. Because the discoloration of a fluorescent sample becomes stronger as the excited light (laser beam) irradiated on the sample gets stronger. Therefore, an observer normally checks the balance of the discoloration of the sample and the acquired image noise and tries to make the amount of excited light as small as possible within the allowable range. For this kind of microscope, therefore, it is very important to suppress the fluorescent loss as much as possible.
We will now discuss a sample marked with two fluorescent dyes (DAPI, CY5) as one example. DAPI has an absorption wavelength in the UV range (340 to 365 nm) and an emitted fluorescent wavelength whose peak appears at approximately 450 nm. CY5 has an absorption wavelength in the red range (630 to 650 nm) and a fluorescent wavelength whose peak appears at approximately 670 nm.
The size of the spot which is formed at the position where those fluorescent rays form an image (where a confocal aperture is provided) is given by the following equation in, for example, Jpn. Pat. Appln. Kokai Publication No. Hei 9-502269.Ø=1.22×λ/NA where NA is the numerical aperture for emission of a lens and λ is the wavelength. The comparison of the spot size of DAPI (fluorescent wavelength of 450 nm) with that of CY5 (fluorescent wavelength of 670 nm), both calculated from the above equation, show that the spot size of CY5 is about 1.5 time greater than that of DAPI.
According to the above-described prior art, therefore, the size of a confocal aperture is set in accordance with the spot size of DAPI in order to secure the confocal effect. This means that the setting of the confocal aperture is set optimized for DAPI, but is too narrow for CY5, resulting in loss of precious fluorescence. Setting the size of the confocal aperture for CY5, on the other hand, would result in an insufficient confocal effect for DAPI.
The bundle of rays that have passed the confocal aperture is resolved by the spectral resolving means (prism) and is split into wavelengths of the individual fluorescent rays using a variable slit. When a prism is used as the spectral resolving means, however, if the size of the bundle of incident rays is large, crossover of the individual wavelengths after spectral resolving occurs, the bundle of rays would not be split into the individual photosensing paths at a sufficient precision.